An estrogen receptor α-derived peptide improves glucose homeostasis during obesity

Abstract

Estrogen receptor α (ERα) plays a crucial role in regulating glucose and energy homeostasis during type 2 diabetes mellitus (T2DM). However, the underlying mechanisms remain incompletely understood. Here we find a ligand-independent effect of ERα on the regulation of glucose homeostasis. Deficiency of ERα in the liver impairs glucose homeostasis in male, female, and ovariectomized (OVX) female mice. Mechanistic studies reveal that ERα promotes hepatic insulin sensitivity by suppressing ubiquitination-induced IRS1 degradation. The ERα 1-280 domain mediates the ligand-independent effect of ERα on insulin sensitivity. Furthermore, we identify a peptide based on ERα 1-280 domain and find that ERα-derived peptide increases IRS1 stability and enhances insulin sensitivity. Importantly, administration of ERα-derived peptide into obese mice significantly improves glucose homeostasis and serum lipid profiles. These findings pave the way for the therapeutic intervention of T2DM by targeting the ligand-independent effect of ERα and indicate that ERα-derived peptide is a potential insulin sensitizer for the treatment of T2DM.

Fig. 1. Hepatic ERα knockout impairs glucose tolerance and insulin sensitivity in both male and female mice.

a IRS1, IRS2, and ERα protein levels in the livers of random-feeding WT and db/db mice, n = 4 mice/group; for IRS1, P = 0.0002; for IRS2, P = 0.0183; for ERα, P = 0.0045. b mRNA expression levels of ERα in the livers of random feeding WT and db/db mice, n = 4 mice/group; P = 0.0078. c mRNA expression levels of ERα in the livers of humans with diabetes, n = 4 (Diabetes poorly controlled) and 5 (Health and Diabetes well controlled); P = 0.0444. d Person correlation coefficient between the Fasting blood glucose/HbA1C/HOMA-IR/Fasting insulin and the mRNA expression levels of ERα in the livers of humans, n = 17 (Fasting blood glucose, Fasting insulin, and HOMA-IR) and 18 (HbA1C). e Random feeding and 5 h fasting blood glucose in ERα^F/F and ERα^LivKO male mice, n = 6 (ERα^LivKO) and 9 (ERα^F/F) mice/group; feeding blood glucose, P = 0.0002; 5 h fasting blood glucose, P = 0.0194. f Glucose tolerance tests in ERα^F/F and ERα^LivKO male mice, n = 8 (ERα^LivKO) and 10 (ERα^F/F) mice/group; 30 min, P = 0.0124; 60 min, P = 0.0013; 90 min, P = 0.0032; 120 min, P = 0.0070. g Insulin tolerance tests in ERα^F/F and ERα^LivKO male mice, n = 6 (ERα^LivKO) and 7 (ERα^F/F) mice/group; 15 min, P = 0.0078. h Insulin signaling was detected in livers from ERα^F/F and ERα^LivKO male mice injected with 2 U insulin for 5 min. The experiments were repeated independently three times. Representative blots were shown. i Random feeding and 5 h fasting blood glucose in ERα^F/F and ERα^LivKO female mice, n = 7 mice/group; feeding blood glucose, P = 0.0011; 5 h fasting blood glucose, P = 0.0299. j Glucose tolerance tests in ERα^F/F and ERα^LivKO female mice, n = 7 mice/group; 30 min, P = 0.0185; 60 min, P = 0.0005; 90 min, P = 0.0444. k Insulin tolerance tests in ERα^F/F and ERα^LivKO female mice, n = 7 mice/group; 30 min, P = 0.0437; 45 min, P = 0.0309. l Insulin signaling was detected in livers from ERα^F/F and ERα^LivKO female mice injected with 2 U insulin for 5 min. The experiments were repeated independently three times. Representative blots were shown. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Two-tailed Student’s t test (a, b, e–g, i–k) or One-way ANOVA with Tukey’s multiple comparisons test (c). Source data are provided as a Source Data file.

Fig. 2. Deletion of hepatic ERα impairs glucose tolerance and insulin sensitivity in both DIO male and female mice.

a Body weight of ERα^F/F and ERα^LivKO male mice treated with HFD, n = 6 (ERα^F/F) and 9 (ERα^LivKO) mice/group. b Body composition of ERα^F/F and ERα^LivKO male mice treated with HFD, n = 6 (ERα^F/F) and 9 (ERα^LivKO) mice/group. c Random feeding and 16 h fasting blood glucose in ERα^F/F and ERα^LivKO male mice treated with HFD, n = 6 (ERα^F/F) and 8 (ERα^LivKO) mice/group; feeding blood glucose, P = 0.0275; 16 h fasting blood glucose, P = 0.0303. d Glucose tolerance tests in ERα^F/F and ERα^LivKO male mice treated with HFD, n = 6 (ERα^F/F) and 8 (ERα^LivKO) mice/group; 0 min, P = 0.0303; 15 min, P = 0.0064; 30 min, P = 0.0008; AUC, P = 0.0149. e Insulin tolerance tests in ERα^F/F and ERα^LivKO male mice treated with HFD, n = 6 (ERα^F/F) and 8 (ERα^LivKO) mice/group; 30 min, P = 0.0461; 60 min, P = 0.0443. f Body weight of ERα^F/F and ERα^LivKO female mice treated with HFD, n = 7 (ERα^LivKO) and 10 (ERα^F/F) mice/group; P = 0.0494. g Body composition of ERα^F/F and ERα^LivKO female mice treated with HFD, n = 7 (ERα^LivKO) and 10 (ERα^F/F) mice/group; fat mass, P = 0.0146; lean mass, P = 0.0115. h Random feeding and 16 h fasting blood glucose in ERα^F/F and ERα^LivKO female mice treated with HFD, n = 7 (ERα^LivKO) and 10 (ERα^F/F) mice/group; random feeding blood glucose, P = 0.0006. i Glucose tolerance tests in ERα^F/F and ERα^LivKO female mice treated with HFD, n = 7 (ERα^LivKO) and 10 (ERα^F/F) mice/group; 30 min, P = 0.0039; 90 min, P = 0.0340; 120 min, P < 0.0001; AUC, P = 0.0036. j Insulin tolerance tests in ERα^F/F and ERα^LivKO female mice treated with HFD, n = 7 (ERα^LivKO) and 10 (ERα^F/F) mice/group; 15 min, P = 0.0088; 45 min, P = 0.0041; 60 min, P = 0.0028. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, unpaired Two-tailed Student’s t test. Source data are provided as a Source Data file.

Fig. 3. Hepatic ERα regulates glucose homeostasis and insulin sensitivity in an IRS1/2-independent manner.

a Random feeding and 16 h fasting blood glucose levels in CNTR, DKO, and TKO male mice under regular chow diet, n = 7 (DKO), 9 (TKO), and 11 (control) mice/group; for feeding blood glucose, CNTR versus DKO, P = 0.0007, CNTR versus TKO, P = 0.0379; for fasting blood glucose, CNTR versus DKO, P = 0.0002, CNTR versus TKO, P < 0.0001. b, c Glucose tolerance tests in in control, DKO, and TKO male mice under regular chow diet, n = 7 (DKO) and 11 (control and TKO) mice/group; CNTR versus DKO, P < 0.0001, CNTR versus TKO, P < 0.0001. d, e Insulin tolerance tests in control, DKO, and TKO male mice under regular chow diet, n = 7 (DKO) and 11 (control and TKO) mice/group; CNTR versus DKO, P = 0.0096, CNTR versus TKO, P = 0.0053. f Random feeding and 16 h fasting blood glucose levels in control, DKO, and TKO female mice under regular chow diet, n = 6 (DKO and TKO) and 8 (control) mice/group; for feeding blood glucose, DKO versus TKO, P = 0.0002; for fasting blood glucose, DKO versus TKO, P < 0.0001. g, h Glucose tolerance tests in in control, DKO, and TKO female mice under regular chow diet, n = 7 (DKO and TKO) and 11 (control) mice/group; CNTR versus DKO, P < 0.0001, CNTR versus TKO, P < 0.0001, DKO versus TKO, P = 0.0024. i, j Insulin tolerance tests in control, DKO, and TKO female mice under regular chow diet, n = 7 (DKO and TKO) and 11 (control) mice/group; CNTR versus DKO, P = 0.0016, CNTR versus TKO, P < 0.0001, DKO versus TKO, P = 0.0452. k Random feeding and 16 h fasting blood glucose levels in control and DKO male/female mice under regular chow diet, n = 5 (Male-DKO and Female-CNTR) and 6 (Male-CNTR and Female-DKO) mice/group; for feeding blood glucose, Male-CNTR versus Male-DKO, P < 0.0001, Male-DKO versus Female-DKO, P < 0.0001; for fasting blood glucose, Male-CNTR versus Male-DKO, P = 0.0095, Male-DKO versus Female-DKO, P = 0.0270. l, m Glucose tolerance tests in control and DKO male/female mice under regular chow diet, n = 5 (Male-DKO and Female-CNTR) and 6 (Male-CNTR and Female-DKO) mice/group; Male-CNTR versus Male-DKO, P < 0.0001, Female-CNTR versus Female-DKO, P < 0.0001, Male-DKO versus Female-DKO, P = 0.0242. n, o Insulin tolerance tests in control and DKO male/female mice under regular chow diet, n = 5 (Male-DKO and Female-CNTR) and 6 (Male-CNTR and Female-DKO) mice/group; Male-CNTR versus Male-DKO, P < 0.0001, Female-CNTR versus Female-DKO, P = 0.0005, Male-CNTR versus Female-CNTR, P = 0.0318, Male-DKO versus Female-DKO, P = 0.0003. p Random feeding and 16 h fasting blood glucose levels in control and DKO male/OVX female mic under regular chow diet, n = 5 (Male-DKO, OVX Female-CNTR, and OVX Female-DKO) and 6 (Male-CNTR) mice/group; for feeding blood glucose, Male-CNTR versus Male-DKO, P = 0.0013, OVX Female-CNTR versus OVX Female-DKO, P = 0.0473, for fasting blood glucose, Male-CNTR versus Male-DKO, P = 0.0004, OVX Female-CNTR versus OVX Female-DKO, P = 0.0032. q, r Glucose tolerance tests in control and DKO male/OVX female mice under regular chow diet, n = 5 (Male-DKO, OVX Female-CNTR, and OVX Female-DKO) and 6 (Male-CNTR) mice/group; Male-CNTR versus Male-DKO, P < 0.0001, OVX Female-CNTR versus OVX Female-DKO, P < 0.0001. s, t Insulin tolerance tests in control and DKO male/OVX female mice under regular chow diet, n = 5 mice/group; Male-CNTR versus Male-DKO, P = 0.0045, OVX Female-CNTR versus OVX Female-DKO, P = 0.0075. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, One-way ANOVA (a–j) or Two-way ANOVA with Tukey’s multiple comparisons test (k–t). CNTR: Control. Source data are provided as a Source Data file.

Fig. 4. ERα increases IRS1 protein stability and promotes hepatic insulin sensitivity.

a Insulin signaling activity in primary hepatocytes from ERα^F/F and ERα^LivKO male mice, n = 3 independent cells; for IRS1, ERα^F/F-Vehicle versus ERα^LivKO-Vehicle, P < 0.0001, ERα^F/F-Insulin versus ERα^LivKO-Insulin, P < 0.0001; for pAKT-S473, ERα^F/F-Vehicle versus ERα^F/F-Insulin, P < 0.0001, ERα^LivKO-Vehicle versus ERα^LivKO-Insulin, P = 0.0186, ERα^F/F-Insulin versus ERα^LivKO-Insulin, P < 0.0001; for ERα, ERα^F/F-Vehicle versus ERα^LivKO-Vehicle, P = 0.0002, ERα^F/F-Insulin versus ERα^LivKO-Insulin, P = 0.0001. b Effect of ERα gain-of-function on hepatic insulin sensitivity. The experiments were repeated independently twice. Representative blots were shown. c Effect of ERα deletion on insulin-induced IRS1 and p85 interaction. The experiments were repeated independently twice. Representative blots were shown. d Effect of ERα gain-of-function on insulin-induced IRS1 and p85 interaction. The experiments were repeated independently twice. Representative blots were shown. e Effect of ERα deletion on mRNA expression of Irs1 and Irs2 in primary mouse hepatocytes, n = 4 (ERα^F/F) and 5 (ERα^LivKO); for ERα, P = 0.0001. f Interaction between ERα and IRS1 or IRS2 in HEK293T cells. The experiments were repeated independently three times. Representative blots were shown. g Effect of ERα on IRS1 ubiquitination in HEK293T cells. The experiments were repeated independently twice. Representative blots were shown. h Effect of ERα on IRS2 ubiquitination in HEK293T cells. The experiments were repeated independently twice. Representative blots were shown. i Effect of ERα deletion on IRS1 ubiquitination in primary mouse hepatocytes. The experiments were repeated independently twice. Representative blots were shown. j Diagram of IRS1 and its truncated domains. k Interaction between ERα and IRS1 domains in HEK293T cells. The experiments were repeated independently three times. Representative blots were shown. l Effect of ERα deletion on IRS1 phosphorylation in primary mouse hepatocytes, n = 3 independent cells; for IRS1, P = 0.0102; for pIRS1-S302, P = 0.0236. m Effect of ERα gain-of-function on palmitate-induced insulin resistance in primary mouse hepatocytes. The experiments were repeated independently twice. Representative blots were shown. Data are presented as mean ± SEM. *P < 0.05, ***P < 0.001, ****P < 0.0001, unpaired Two-tailed Student’s t test (e and l), Two-way ANOVA with Tukey’s multiple comparisons test (a). Source data are provided as a Source Data file.

Fig. 5. ERα 1-280 domain increases IRS1 protein stability and enhances insulin signaling.

a Diagram of truncated ERα protein domains. b Effect of ERα on insulin sensitivity in HepG2 cells, n = 3 independent cells; for IRS1, Vehicle versus ERα-Vehicle, P < 0.0001, Insulin versus ERα-Insulin, P = 0.0005; for pAKT-S473, Vehicle versus ERα-Vehicle, P = 0.0025, Vehicle versus Insulin, P < 0.0001, ERα-Vehicle versus ERα-Insulin, P < 0.0001, Insulin versus ERα-Insulin, P = 0.0001; for pAKT-T308, Vehicle versus Insulin, P < 0.0001, ERα-Vehicle versus ERα-Insulin, P < 0.0001, Insulin versus ERα-Insulin, P < 0.0001. c Effect of ERα 1-280 on insulin sensitivity in HepG2 cells, n = 3 independent cells; for IRS1, Vehicle versus ERα 1-280-Vehicle, P < 0.0001, Insulin versus ERα 1-280-Insulin, P = 0.0034; for pAKT-S473, Vehicle versus ERα 1-280-Vehicle, P = 0.0457, Vehicle versus Insulin, P < 0.0001, ERα 1-280-Vehicle versus ERα 1-280-Insulin, P < 0.0001, Insulin versus ERα 1-280-Insulin, P = 0.0007; for pAKT-T308, Vehicle versus Insulin, P = 0.0040, ERα 1-280-Vehicle versus ERα 1-280-Insulin, P = 0.0003, Insulin versus ERα 1-280-Insulin, P = 0.0397. d Effect of ERα DBD + AF2 on insulin sensitivity in HepG2 cells, n = 3 independent cells; for pAKT-S473, Vehicle versus Insulin, P < 0.0001, ERα DBD + AF2-Vehicle versus ERα DBD + AF2-Insulin, P < 0.0001; for pAKT-T308, Vehicle versus Insulin, P = 0.0002, ERα DBD + AF2-Vehicle versus ERα DBD + AF2-Insulin, P = 0.0003. e Interaction between ERα 1-280 and IRS1 protein domains. The experiments were repeated independently twice. Representative blots were shown. f Effect of ERα 1-280 on IRS1 phosphorylation at S302 in HepG2 cells, n = 3 independent cells; for pIRS1-S302, P = 0.0227; for IRS1, P = 0.0187. g Effect of ERα on IRS1 phosphorylation at S302 in HepG2 cells, n = 3 independent cells; for pIRS1-S302, P = 0.0207; for IRS1, P = 0.0154. h Effect of ERα 1-280 on IRS1 ubiquitination in HEK293T cells. The experiments were repeated independently twice. Representative blots were shown. i Effect of ERα 1-280 on IRS2 ubiquitination in HEK293T cells. The experiments were repeated independently twice. Representative blots were shown. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, unpaired Two-tailed Student’s t test (f, g), Two-way ANOVA with Tukey’s multiple comparisons test (b, c, d). CNTR: Control. Source data are provided as a Source Data file.

Fig. 6. ERα-derived AF1 peptide increases hepatic insulin sensitivity through IRS1.

a SVM score of interaction between IRS1 and ERα domains. b Effect of ERα 1-60 domain and AF1 peptide (34 aa) on insulin sensitivity in HepG2 cells, n = 3 independent cells; for IRS1, Insulin versus 34 aa peptide-Insulin, P = 0.0046, Insulin versus ERα 1-60-Insulin, P = 0.0006; for pAKT-S473, Vehicle versus Insulin, P < 0.0001, Insulin versus 34 aa peptide-Insulin, P = 0.0044, Insulin versus ERα 1-60-Insulin, P = 0.0030. c Effect of ERα 60-280 domain on insulin sensitivity in HepG2 cells, n = 3 independent cells; for pAKT-S473, Vehicle versus Insulin, P < 0.0001, ERα 60-280-Vehicle versus ERα 60-280-Insulin, P < 0.0001. d Immunofluorescence staining of AF1 peptide (34 aa) and ERα 1-60 domain in HEK293T cells. The experiments were repeated independently twice. Representative images were shown. e Amino acid sequence of FITC labeled AF1 peptide conjugated with TAT. f Interaction between AF1 peptide and IRS1 protein domains. The experiments were repeated independently twice. Representative results were shown. g Effect of AF1 peptide on IRS1 phosphorylation at S302 in primary mouse hepatocytes, n = 3 independent cells; for pIRS1-S302, P = 0.0377; for IRS1, P = 0.0156. h Effect of AF1 peptide on IRS1 ubiquitination in HEK293T cells. The experiments were repeated independently twice. Representative blots were shown. i Effect of AF1 peptide on insulin-induced suppression of HGP in primary hepatocytes upon glucagon treatment, n = 4 independent cells; CNTR-Vehicle versus CNTR-Glucagon, P = 0.0006, AF1 peptide-Vehicle versus AF1 peptide-Glucagon, P = 0.0272, CNTR-Glucagon-Insulin versus AF1 peptide-Glucagon-Insulin, P = 0.0019. j Effect of AF1 peptide on insulin sensitivity in primary mouse hepatocytes, n = 3 independent cells; for IRS1, CNTR-Vehicle versus AF1 peptide-Vehicle, P = 0.0067, CNTR-Insulin versus AF1 peptide-Insulin, P = 0.0314, for pAKT-S473, AF1 peptide-Vehicle versus AF1 peptide-Insulin, P = 0.0001, CNTR-Insulin versus AF1 peptide-Insulin, P = 0.0008; for pAKT-T308, CNTR-Vehicle versus CNTR-Insulin, P = 0.0154, AF1 peptide-Vehicle versus AF1 peptide-Insulin, P = 0.0002, CNTR-Insulin versus AF1 peptide-Insulin, P = 0.0167. k Effect of AF1 peptide on insulin sensitivity in control and ERα deficient primary mouse hepatocytes, n = 3 independent cells, for IRS1, ERα^F/F-Insulin versus ERα^LivKO-Insulin, P < 0.0001, ERα^LivKO-Insulin versus ERα^LivKO-Insuin-AF1 peptide, P = 0.0061, for pAKT-S473, ERα^F/F-Vehicle versus ERα^F/F-Insulin, P < 0.0001, ERα^F/F-Insulin versus ERα^LivKO-Insulin, P < 0.0001, ERα^LivKO-Insulin versus ERα^LivKO-Insulin-AF1 peptide, P = 0.0034; for pAKT-T308, ERα^F/F-Vehicle versus ERα^F/F-Insulin, P < 0.0001, ERα^F/F-Insulin versus ERα^LivKO-Insulin, P < 0.0001, ERα^LivKO-Insulin versus ERα^LivKO-Insulin-AF1 peptide, P = 0.0090. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, unpaired Two-tailed Student’s t test (g), One-way ANOVA (b, i, k), Two-way ANOVA with Tukey’s multiple comparisons test (c, j). CNTR: Control. Source data are provided as a Source Data file.

Fig. 7. AF1 peptide improves glucose tolerance and insulin sensitivity in diabetic mice.

a Body weight of db/db mice treated with control and AF1 peptide for 5 weeks, n = 5 (AF1 peptide) and 6 (control) mice/group. b Random feeding and 16 h fasting blood glucose in db/db mice treated with control and AF1 peptide for 5 weeks, n = 6 mice/group; fasting blood glucose, P = 0.0409; feeding blood glucose, P = 0.0317. c Glucose tolerance tests in db/db mice treated with control and AF1 peptide for 5 weeks, n = 6 mice/group; 0 min, P = 0.0409; 15 min, P = 0.0006; 30 min, P = 0.0032; 90 min, P = 0.0375; 120 min, P = 0.0104; AUC, P = 0.0042. d Insulin tolerance tests in db/db mice treated with control and AF1 peptide for 5 weeks, n = 6 mice/group; 30 min, P = 0.0328; 45 min, P = 0.0198, AUC, P = 0.0217. e H&E staining of livers from db/db mice treated with control and AF1 peptide. Scale: 200 µm. Liver fat content was calculated, n = 5 mice/group; P = 0.0013. Representative images were shown. f Liver insulin sensitivity in db/db mice treated with control and AF1 peptide, n = 5 mice/group; pIRS1-S302, P < 0.0001, IRS1, P = 0.0040, pAKT-S473, P = 0.0026, pAKT-T308, P = 0.0037. g Phosphorylation of AKT in epididymal fat and skeleton muscle from db/db mice treated with control and AF1 peptide, n = 4 mice/group; for fat, pAKT-S473, P = 0.0257, pAKT-T308, P = 0.0491, for muscle, pAKT-S473, P = 0.0069, pAKT-T308, P = 0.0450. h Volcano plots of differentially expressed genes (DEGs) in livers from db/db mice treated with control and AF1 peptide. Genes upregulated or downregulated by more than 1.3-fold are shown in red and blue, respectively. i KEGG pathway analysis of DEGs in livers from db/db mice treated with control and AF1 peptide. j Heatmap of representative DEGs in livers from db/db mice treated with control and AF1 peptide. k Liver triglyceride, serum triglyceride, AST, cholesterol, LDL, and NEFA levels in db/db mice treated with control and AF1 peptide, n = 4 (serum LDL of CNTR group), 5 (liver triglyceride and serum cholesterol of CNTR group as well as serum LDL and NEFA of AF1 peptide group), and 6 (serum triglyceride and AST of CNTR and AF1 peptide group, liver triglyceride and serum cholesterol of AF1 peptide group, and serum NEFAof CNTR group); serum AST, P = 0.0487, serum cholesterol, P = 0.0234, serum LDL, P = 0.0037, serum NEFA, P = 0.0411. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, unpaired Two-tailed Student’s t test. CNTR: Control. Source data are provided as a Source Data file.